UCLA

Drone-based applications are increasingly prevalent in surveillance, delivery, automation, and many more usage scenarios. They all require seamless mobility support and high-throughput data access. The mmWave communication in current cellular networks offers the physical-layer technology candidate, but cannot deliver the required services. The root cause lies in the fundamental limit of the legacy time-frequency domain design, where time-frequency-based multiplexing amplifies Doppler effects and interference in the drone setting, thus incurring packet losses and connectivity failures.

This dissertation tackles the problem from a novel cross-domain approach to embracing mmWave for 6G and beyond cellular networks. The cross-domain design synthesizes the delay-Doppler domain and time-frequency domain. The insight is that we may achieve the best of both worlds if properly multiplexing signaling in both domains. Along this new direction, we seek to achieve three goals: high-fidelity channel estimation under aerial mobility, effective interference cancellation without pre-transmission negotiation, and readily-deployable practical solutions in 6G and beyond.

Our contributions can be summarized as follows: 1) We design a novel concurrent channel state inference algorithm for drones by modeling the shared multipathing features between cells. Our key insight is that the delay-Doppler domain reveals frequency-decoupled physical path features. We devise the delay-Doppler signaling that is embedded into the legacy OFDM signals to enable Doppler-independent inference. Our design characterizes and decouples path delay, and infers multipath propagation. 2) We design a novel cross-domain interference cancellation scheme by multiplexing signals and data across neighboring cells. Unlike prior schemes that require pre-transmission negotiation of power and data rate, we transform transmission-induced interference between cells into meaningful data, boosting throughput among dense, interfering cells in the sky. Our solution is resilient to multiple interferer scenarios under dense cell deployment. 3) We integrate cross-domain signals in 6G and beyond with practical system designs. Our design has one key objective: retain the current data transmission pipeline within 3GPP standardized framework to maximize the deployability. We devise an overlay to embed new delay-Doppler signals into the current physical layer in the time-frequency domain. We then propose adaptive scheduling so that cross-domain signals can be readily integrated into current standards. We assess the feasibility of our approaches by prototyping and evaluating with an mmWave testbed.